How Many Bits Does It Take To Kill You? 300
pegr writes "Andrew 'bunnie' Huang, Reverse Engineer, XBox hacker, and generally smart guy, muses over the H1N1/swine flu virus as only a reverse engineer can: 'I now know how to modify the virus sequence to probably make it more deadly.' Not that he would, of course. bunnie has consistently made the esoteric available to us mere mortals, and his overview of the H1N1 virus is a fascinating read from a unique perspective." (Seen today also at the top of Schneier on Security.)
Re:fascinating! (Score:5, Interesting)
(Replying to my own comment.)
That said, it's a quite well-written tutorial-style article with engaging prose that tackles a number of the relevant issues. I just balked at the "reverse engineer takes on biology" angle, as if that were something biologists had never thought of.
Re:fascinating! (Score:5, Interesting)
If only biologists had thought of the idea of treating DNA/RNA sequences as data, and then analyzing their properties statistically and computationally, with an eye towards what effects different modifications to the sequences might be predicted to have. We might call this field something fancy like "biological informatics".
Hahaha, I'm sure the biological informaticians are laughing their asses off. Kinda like we computer geeks did when the Not So Hon. Ted Stevens described the Internet as a "series of tubes".
Meanwhile, though, I'm really enjoying the analogies that "bunnie" draws between DNA/RNA and computer bits. You see, I know a thing or two about computer bits, and ports, and stuff like that. And I know that DNA encodes proteins. But I didn't make the connection the way "bunnie" does, with a simple statement like this:
That's probably baby science to a biological informatician, just like mapping to port 25 is baby networking to many of us. But for me, it makes the concepts click.
Similarly, we all made fun of the "series of tubes" metaphor, without considering that for most of humanity, an electron is "the size and shape of a small pea" (Heinlein reference). If thinking of the Internet as a bunch of interconnected steampunk-style tubes that can get full (saturated bandwidth) helps a non-techie understand why they can't watch YouTube and play Halo at the same time... well, so much the better.
Re:fascinating! (Score:1, Interesting)
Re:Increasing mortality is bad for business (Score:5, Interesting)
Actually, HIV has become less deadly as time goes by. There's been selective pressure for it to kill the hosts less slowly:
http://health.dailynewscentral.com/content/view/1716/ [dailynewscentral.com]
Re:Increasing mortality is bad for business (Score:3, Interesting)
> The second-most successful virus was the one that struck the
> Roman Empire circa 600 A.D. because if that virus had not struck,
> the Eastern Roman Emperor's army would have succeeded in his
> mission to reclaim Italy, Rome, and possibly France/Gaul too.
Was this after Narses the court eunuch and general conquered Italy, then let the Lombards attack the North to show the Emperor that he was needed (and committed suicide in shame when they succeeded)?
> he most-successful virus struck Europe in the mid-1200s,
> Thus the middle class was born.
The middle class existed for long before that. It merely improved the lot of peasants for about 60 years (until population levels came back) and created the "Rotten Boroughs" in England (abandoned towns that didn't lose their representation in Commons until the early 19th Century).
Anyway, the common cold beats them in almost any two year period. Further, people continue to catch colds all through their lives.
Now, if the goal of the virus were to wipe out humanity or at least change history, then your viruses would have won. Prove that either was deliberately weaponized, or introduced by aliens making a multi-sense recording for "viewers" in the Galactic Community (to make a season-ending cliffhanger, or else because a new bunch of writers wanted to "reboot" the franchise), and I will accept your definition of successful.
Re:Port H1 maps specifically to the windpipe regio (Score:3, Interesting)
Only if the firewall also performs deep packet inspection. Many bad critters (viruses/bacteria) enter the system by making our firewall(s) think they are innocuous by externally looking link other good critters. It is the payload that is the real problem. If we could teach the body to somehow read the payload before docking with the receptors we could be disease (contracted from viruses/bacteria) free.
Nanoprobe-supported organs. Once again, Star Trek has beaten us to it.
more bioinformatics for beginners (Score:5, Interesting)
Re:It's humbling that I could be killed by 3.2kbyt (Score:5, Interesting)
It would actually take less than that, though it wouldn't spread the same way. Remember that prions are proteins that can kill you rather than whole viruses. The protein that gets misfolded in Bovine Spongiform Encephalopathy (or mad cow) seems to be called just Prion protein [nih.gov] and is only 253 amino acids. If bunnie is correct and one amino acid = 6 bits, then thats 1,518 bits. "Bit calculator" [matisse.net] tells me that would be 0.185 kbytes.
Granted, this wouldn't be airborne death, would be extremely slow, and wouldn't cause a pandemic, but still, far less data.
Even if you were to go the viral route, at least one virus is tricky in that it produces multiple proteins from overlapping reading frames. [nih.gov] That is, the same sections of RNA genome (sendai uses RNA instead of DNA) is read in multiple ways to make different functional proteins, one protein might be formed from reading AUG GAU GGG CAG, which would make the amino acid sequence MDGQ, but that could aso be read as A UGG *AUG* GGC AG where the starred AUG is the start, making a protein of MG. I find that pretty cool, because as Carl Sagan pointed out, try doing that with english. "Romancement to get her" can be spaced differently to produce "roman cement together" is the longest he could come up with and it doesn't even make sense. Viruses make whole proteins that work. Anyway, the point of all that was that viruses can in some cases double up, so it would take even fewer nucleotides to produce the same amount of protiens.
Another interesting observation (Score:5, Interesting)
Looking at the amino acid and codon table [algoart.com] I noticed another interesting point: The triples which code for the same amino acid typically differ only in the last base. Indeed, this can be made stronger: Except for the STOP codon, in each set of codons with no more than four members, the first two bases are always the same (for those with more than four codons that's of course not possible). Moreover, quite a few amino acids have exactly four codons which differ only in the last base, i.e. the amino acid is completely and unambiguously determined by the first two bases alone. Indeed, one can rearrange this into the following 16-entry table:
Note how many lines only have one entry on the right hand side. Could this mean the genetic code evolved from a two-base version (with only 15 amino acids) to the current three-base version?
Re:It's humbling that I could be killed by 3.2kbyt (Score:1, Interesting)
The protein that gets misfolded in Bovine Spongiform Encephalopathy (or mad cow) seems to be called just Prion protein [nih.gov] and is only 253 amino acids. If bunnie is correct and one amino acid = 6 bits, then thats 1,518 bits.
But the same sequence of amino acids exists normally in us, as a useful protein. What makes the prion variant dangerous is that the amino acid chain has folded slightly differently, to form a different shape - so it's in the file metadata, not the file itself.
Re:fascinating! (Score:5, Interesting)
I just balked at the "reverse engineer takes on biology" angle, as if that were something biologists had never thought of.
Interesting that you should say that - the traditional biologists, by and large, don't think of doing things like this. Bioinformatics is a catch-all for any number of different disciplines, all in relative infancy, and almost always pioneered by people outside the traditional biology arenas.
I studied biochemistry in college, with a ton of extra math, physics, and computer science. Then I did a PhD developing DNA diagnostics for flu (awarded by the chem department, but I was a full time programmer and part time bench chemist).
My first paper was applying Shannon informational entropy theory to big alignments of flu DNA to look for conserved regions. No one around me had a clue what the hell I was on about. The code I wrote for that paper is still used by the Flu Division at CDC.
The only place where this article went wrong was in assuming that traits are trivially mapped to sequences. In practice, it almost always turns out to be extremely non-trivial, and in flu it almost doesn't work at all (the biologist figured out the easy cases years ago). Never the less, most really good science starts with some assumption that looks to be extremely over-simplified, and turns out to be very predictive.
There is going to be a lot of room for hackers and coders in the biological sciences in coming years - computer science has solutions to problems the traditional biologists haven't even realized are problems yet. Data storage and retrieval to support high-throughput sequencing labs, new algorithms for large-scale data analysis, instrument networking for lab automation. The job postings will go up just as soon as the biologists figure out that they have a problem...